Sir

Recent analyses have compared the scientific output of countries according to the number of papers published and their citations1,2. These show that the United Kingdom scores highly in many areas.

Although there are reservations about the use of citations, they appear to be broadly accepted for assessment exercises and they no doubt influence policy decisions. Thus there is interest in establishing which sectors of the research base contribute to this success. For example, May1 has speculated that the reason for the United Kingdom's relative success is because a higher proportion of its research is done in universities, although the opposite view has also been expressed3.

Previously published studies do not separate out the contribution of research council institutes to the United Kingdom's total research output. Institutes mostly concentrate on biological and biomedical research, areas in which the United Kingdom was identified1,2 as being strong, and much less on areas such as chemistry, in which the United Kingdom is not in the top five in impact1.

To establish the contribution of a group of 15 institutes sponsored by, or related to, the Biotechnology and Biological Sciences Research Council (BBSRC) and listed in Table 1, the Science Policy Research Unit (SPRU) at the University of Sussex surveyed publications in all 154 fields defined by the independent Institute for Scientific Information (ISI)4. The publications in the SPRU database are derived from that used by ISI5 but include only refereed papers, notes and reviews. SPRU has verified thatapers are allocated to their correct institution6,7.

Table 1 Relative performance of sections of the UK science base in citation impact
Full size table

The results in Table 1 compare the figures from the institutes with those from the United Kingdom as a whole, the total university sector, and a group of 15 universities that did best in fields relevant to the institutes in a recent ISI analysis5.

The relative citation impact (RCI) for the papers in the fields in which the institutes publish most has been calculated for the different groups. In 14 out of the 22 fields, in which the institutes published more than 300 papers, the RCI for the institute papers is higher than for those from any other group. These results remain broadly similar when the total citations from 1981 to 1994 are used in a three- or five-year window of citations.

To estimate international competitiveness, SPRU compared the citation rates for the same fields, using a three-year window, with those for the world, the United States, Canada, Europe as a whole, Germany and France. In all but three fields, the institute average citation rate of papers published between 1981 and 1992 was greater than those from any other country or group of countries.

Relative output from different organizations is difficult to assess because of their varying roles. For example, the papers in Table 1 represent less than half of the total publications of the institutes. But Krebs8 calculated that the non-medical university sector produced 8.26 papers per researcher, compared with 8.43 for institutes; and other sectors had lower figures.

In conclusion, the results presented here suggest that the quality of the refereed papers from the selected institutes are competitive with those from other sectors of the UK research base and with those of other countries. Other studies6,7 (J. S. Katz et al. in preparation) suggest that this conclusion can be extended to other institutes and similar organizations. So the scientific output of the United Kingdom, as measured by citations, is enhanced by papers from research institutes. The institute organization and environment does not inhibit the production of high-impact research.

This letter was prepared with the help of the directors of the institutes in Table 1. The table was produced with the assistance of Dr Sylvan Katz and his colleagues at SPRU. I am grateful to them all for their contributions.